1 Heatmap

我们在分析了差异表达数据之后，经常要生成一种直观图－－热图(heatmap)。这一节就以基因芯片数据为例，示例生成高品质的热图。 heatmap

1.1 Simple

首先还是从最简单的heatmap开始。

library(ggplot2)
library(ALL) #可以使用biocLite("ALL")安装该数据包
## Loading required package: Biobase
## Loading required package: BiocGenerics
## Loading required package: parallel
## 
## Attaching package: 'BiocGenerics'
## The following objects are masked from 'package:parallel':
## 
##     clusterApply, clusterApplyLB, clusterCall, clusterEvalQ,
##     clusterExport, clusterMap, parApply, parCapply, parLapply,
##     parLapplyLB, parRapply, parSapply, parSapplyLB
## The following objects are masked from 'package:stats':
## 
##     IQR, mad, xtabs
## The following objects are masked from 'package:base':
## 
##     anyDuplicated, append, as.data.frame, as.vector, cbind,
##     colnames, do.call, duplicated, eval, evalq, Filter, Find, get,
##     grep, grepl, intersect, is.unsorted, lapply, lengths, Map,
##     mapply, match, mget, order, paste, pmax, pmax.int, pmin,
##     pmin.int, Position, rank, rbind, Reduce, rownames, sapply,
##     setdiff, sort, table, tapply, union, unique, unlist, unsplit
## Welcome to Bioconductor
## 
##     Vignettes contain introductory material; view with
##     'browseVignettes()'. To cite Bioconductor, see
##     'citation("Biobase")', and for packages 'citation("pkgname")'.
data("ALL")
library(limma)
## 
## Attaching package: 'limma'
## The following object is masked from 'package:BiocGenerics':
## 
##     plotMA
eset<-ALL[,ALL$mol.biol %in% c("BCR/ABL","ALL1/AF4")]
f<-factor(as.character(eset$mol.biol))
design<-model.matrix(~f)
fit<-eBayes(lmFit(eset,design)) #对基因芯片数据进行分析，得到差异表达的数据
selected  <- p.adjust(fit$p.value[, 2]) <0.001 
esetSel <- eset[selected,] #选择其中一部分绘制热图
dim(esetSel) #从这尺度上看，数目并不多，但也不少。如果基因数过多，可以分两次做图。
## Features  Samples 
##       84       47
library(hgu95av2.db) #biocLite('hgu95av2.db')
## Loading required package: AnnotationDbi
## Loading required package: stats4
## Loading required package: IRanges
## Loading required package: S4Vectors
## Loading required package: org.Hs.eg.db
## Loading required package: DBI
## 
## 
data<-exprs(esetSel)
probes<-rownames(data)
symbol<-mget(probes,hgu95av2SYMBOL,ifnotfound=NA)
symbol<-do.call(rbind,symbol)
symbol[is.na(symbol[,1]),1]<-rownames(symbol)[is.na(symbol[,1])]
rownames(data)<-symbol[probes,1] #给每行以基因名替换探针名命名，在绘制热图时直接显示基因名。
heatmap(data,cexRow=0.5)

这个图有三个部分，样品分枝树图和基因分枝树图，以及热图本身。之所以对样品进行聚类分析排序，是因为这次的样品本身并没有分组。如果有分组的话，那么可以关闭对样品的聚类分析。对基因进行聚类分析排序，主要是为了色块好看，其实可以选择不排序，或者使用GO聚类分析排序。上面的这种热图，方便简单，效果非常不错。 接下来我们假设样品是分好组的，那么我们想用不同的颜色来把样品组标记出来，那么我们可以使用ColSideColors参数来实现。同时，我们希望变更热图的渐变填充色，可以使用col参数来实现。

color.map <- function(mol.biol) { if (mol.biol=="ALL1/AF4") "#FF0000" else "#0000FF" }
patientcolors <- unlist(lapply(esetSel$mol.bio, color.map))
heatmap(data, col=topo.colors(100), ColSideColors=patientcolors, cexRow=0.5)

在heatmap函数中，样品分组只能有一种，如果样品分组有多次分组怎么办？heatmap.plus就是来解决这个问题的。它们的参数都一致，除了ColSideColors和RowSideColors。heatmap使用是一维数组，而heatmap.plus使用的是字符矩阵来设置这两个参数。

1.2 heatmap.plus

library(heatmap.plus)
hc<-hclust(dist(t(data)))
dd.col<-as.dendrogram(hc)
groups <- cutree(hc,k=5)
color.map <- function(mol.biol) { if (mol.biol=="ALL1/AF4") 1 else 2 }
patientcolors <- unlist(lapply(esetSel$mol.bio, color.map))
col.patientcol<-rbind(groups,patientcolors)
mode(col.patientcol)<-"character"
heatmap.plus(data,ColSideColors=t(col.patientcol),cexRow=0.5)

这样绘图的不足是没有热图色key值。gplots中的heatmap.2为我们解决了这个问题。而且它带来了更多的预设填充色。下面就是几个例子。

1.3 gplots

library("gplots")
## 
## Attaching package: 'gplots'
## The following object is masked from 'package:IRanges':
## 
##     space
## The following object is masked from 'package:stats':
## 
##     lowess
heatmap.2(data, col=redgreen(75), scale="row",ColSideColors=as.character(patientcolors),key=TRUE, symkey=FALSE, density.info="none", trace="none", cexRow=0.5)

heatmap.2(data, col=heat.colors(100), scale="row", ColSideColors=as.character(patientcolors),key=TRUE, symkey=FALSE, density.info="none", trace="none", cexRow=0.5)

heatmap.2(data, col=terrain.colors(100), scale="row", ColSideColors=as.character(patientcolors),key=TRUE, symkey=FALSE, density.info="none", trace="none", cexRow=0.5)

heatmap.2(data, col=cm.colors(100), scale="row", ColSideColors=as.character(patientcolors),key=TRUE, symkey=FALSE, density.info="none", trace="none", cexRow=0.5)

heatmap.2(data, col=redblue(100), scale="row", ColSideColors=as.character(patientcolors),key=TRUE, symkey=FALSE, density.info="none", trace="none", cexRow=0.5)

heatmap.2(data, col=colorpanel(100,low="white",high="steelblue"), scale="row", ColSideColors=as.character(patientcolors),key=TRUE,  symkey=FALSE, density.info="none", trace="none", cexRow=0.5)

1.4 ggplot2

然而，以上的heatmap以及heatmap.2虽然方便简单，效果也很不错，可以使用colorpanel方便的设置渐变填充色，但是它的布局没有办法改变，生成的效果图显得有点呆板，不简洁。为此这里介绍如何使用ggplot2当中的geom_tile来为基因芯片绘制理想的热图。

library(ggplot2)
library(plotrix)
## 
## Attaching package: 'plotrix'
## The following object is masked from 'package:gplots':
## 
##     plotCI
library(scales)
## 
## Attaching package: 'scales'
## The following object is masked from 'package:plotrix':
## 
##     rescale
library(reshape)
## 
## Attaching package: 'reshape'
## The following object is masked from 'package:IRanges':
## 
##     expand
## The following object is masked from 'package:S4Vectors':
## 
##     rename
hc<-hclust(dist(data))
rowInd<-hc$order
hc<-hclust(dist(t(data)))
colInd<-hc$order
data.m<-data[rowInd,colInd] #聚类分析的作用是为了色块集中，显示效果好。如果本身就对样品有分组，基因有排序，就可以跳过这一步。
data.m<-apply(data.m,1,rescale) 
data.m<-t(data.m) #变换以后转置了。
coln<-colnames(data.m) 
rown<-rownames(data.m) #保存样品及基因名称。因为geom_tile会对它们按坐标重排，所以需要使用数字把它们的序列固定下来。
colnames(data.m)<-1:ncol(data.m)
rownames(data.m)<-1:nrow(data.m)
data.m<-melt(data.m) #转换数据成适合geom_tile使用的形式
head(data.m)
##   X1 X2     value
## 1  1  1 0.1898007
## 2  2  1 0.6627467
## 3  3  1 0.5417057
## 4  4  1 0.4877054
## 5  5  1 0.5096474
## 6  6  1 0.2626248
base_size<-12 #设置默认字体大小，依照样品或者基因的多少而微变。
(p <- ggplot(data.m, aes(X2, X1)) + geom_tile(aes(fill = value),colour = "white") + scale_fill_gradient(low = "white", high = "steelblue"))

p + theme_grey(base_size = base_size) + labs(x = "", y = "") + scale_x_continuous(expand = c(0, 0), labels=coln, breaks=1:length(coln)) + scale_y_continuous(expand = c(0, 0),labels=rown,breaks=1:length(rown)) + theme(
    axis.ticks = element_blank(),
    axis.text.x = element_text(size = base_size * 0.8, angle = 90, hjust = 0, colour = "grey50"),
    axis.text.y = element_text(size = base_size * 0.5, hjust=1, colour="grey50"))

(p <- ggplot(data.m, aes(X2, X1)) + geom_tile(aes(fill = value),colour = "white") + scale_fill_gradient(low = "yellow",high = "red"))

p + theme_grey(base_size = base_size) + labs(x = "",y = "") + scale_x_continuous(expand = c(0, 0),labels=coln,breaks=1:length(coln)) +
scale_y_continuous(expand = c(0, 0),labels=rown,breaks=1:length(rown)) + theme(
     axis.ticks = element_blank(), 
     axis.text.x = element_text(size = base_size * 0.8, angle = 90, hjust = 0, colour = "grey50"), 
     axis.text.y = element_text(size = base_size * 0.8, hjust=1, colour="grey50"))

(p <- ggplot(data.m, aes(X2, X1)) + geom_tile(aes(fill = value),
     colour = "white") + scale_fill_gradient(low = "green",
     high = "red"))

p + theme_grey(base_size = base_size) + labs(x = "",
     y = "") + scale_x_continuous(expand = c(0, 0),labels=coln,breaks=1:length(coln)) +
     scale_y_continuous(expand = c(0, 0),labels=rown,breaks=1:length(rown)) + theme(
     axis.ticks = element_blank(), axis.text.x = element_text(size = base_size *
     0.8, angle = 90, hjust = 0, colour = "grey50"), axis.text.y = element_text(
     size = base_size * 0.8, hjust=1, colour="grey50"))

(p <- ggplot(data.m, aes(X2, X1)) + geom_tile(aes(fill = value),
     colour = "white") + scale_fill_gradient(low = "seagreen",
     high = "white"))

p + theme_grey(base_size = base_size) + labs(x = "",
     y = "") + scale_x_continuous(expand = c(0, 0),labels=coln,breaks=1:length(coln)) +
     scale_y_continuous(expand = c(0, 0),labels=rown,breaks=1:length(rown)) + theme(
     axis.ticks = element_blank(), axis.text.x = element_text(size = base_size *
     0.8, angle = 90, hjust = 0, colour = "grey50"), axis.text.y = element_text(
     size = base_size * 0.8, hjust=1, colour="grey50"))

(p <- ggplot(data.m, aes(X2, X1)) + geom_tile(aes(fill = value),
     colour = "white") + scale_fill_gradient(low = "white",
     high = "sienna4"))

p + theme_grey(base_size = base_size) + labs(x = "",
     y = "") + scale_x_continuous(expand = c(0, 0),labels=coln,breaks=1:length(coln)) +
     scale_y_continuous(expand = c(0, 0),labels=rown,breaks=1:length(rown)) + theme(
     axis.ticks = element_blank(), axis.text.x = element_text(size = base_size *
     0.8, angle = 90, hjust = 0, colour = "grey50"), axis.text.y = element_text(
     size = base_size * 0.8, hjust=1, colour="grey50"))

(p <- ggplot(data.m, aes(X2, X1)) + geom_tile(aes(fill = value),
     colour = "white") + scale_fill_gradient(low = "black",
     high = "gray85"))

p + theme_grey(base_size = base_size) + labs(x = "",
     y = "") + scale_x_continuous(expand = c(0, 0),labels=coln,breaks=1:length(coln)) +
     scale_y_continuous(expand = c(0, 0),labels=rown,breaks=1:length(rown)) + theme(
     axis.ticks = element_blank(), axis.text.x = element_text(size = base_size *
     0.8, angle = 90, hjust = 0, colour = "grey50"), axis.text.y = element_text(
     size = base_size * 0.8, hjust=1, colour="grey50"))

1.5 Lattice

除了ggplot2，还有lattice也是不错的选择。我只使用一种填充色，生成两个图，以作示例。

hc<-hclust(dist(data))
dd.row<-as.dendrogram(hc)
row.ord<-order.dendrogram(dd.row) #介绍另一种获得排序的办法
hc<-hclust(dist(t(data)))
dd.col<-as.dendrogram(hc)
col.ord<-order.dendrogram(dd.col)
data.m<-data[row.ord,col.ord]
library(ggplot2)
data.m<-apply(data.m,1,rescale) #rescale是ggplot2当中的一个函数
library(lattice)
levelplot(data.m,
          aspect = "fill",xlab="",ylab="",
          scales = list(x = list(rot = 90, cex=0.8),y=list(cex=0.5)),
          colorkey = list(space = "left"),col.regions = heat.colors)

library(latticeExtra)
## Loading required package: RColorBrewer
## 
## Attaching package: 'latticeExtra'
## The following object is masked from 'package:ggplot2':
## 
##     layer
levelplot(data.m,
          aspect = "fill",xlab="",ylab="",
          scales = list(x = list(rot = 90, cex=0.5),y=list(cex=0.4)),
          colorkey = list(space = "left"),col.regions = heat.colors,
          legend =
          list(right =
               list(fun = dendrogramGrob, #dendrogramGrob是latticeExtra中绘制树型图的一个函数
                    args =
                    list(x = dd.row, ord = row.ord,
                         side = "right",
                         size = 5)),
               top =
               list(fun = dendrogramGrob,
                    args =
                    list(x = dd.col, 
                         side = "top",
                         type = "triangle")))) #使用三角型构图

1.6 Pretty Heat Maps

可是可是，绘制一个漂亮的热图这么难么？参数如此之多，设置如此复杂，色彩还需要自己指定。有没有简单到发指的函数呢？有！那就是pheatmap，全称pretty heatmaps.

library(pheatmap)
pheatmap(data,fontsize=9, fontsize_row=6) #最简单地直接出图

pheatmap(data, scale = "row", clustering_distance_row = "correlation", fontsize=9, fontsize_row=6) #改变排序算法

pheatmap(data, color = colorRampPalette(c("navy", "white", "firebrick3"))(50), fontsize=9, fontsize_row=6) #自定义颜色

pheatmap(data, cluster_row=FALSE, fontsize=9, fontsize_row=6) #关闭按行排序

pheatmap(data, legend = FALSE, fontsize=9, fontsize_row=6) #关闭图例

pheatmap(data, cellwidth = 6, cellheight = 5, fontsize=9, fontsize_row=6) #设定格子的尺寸

color.map <- function(mol.biol) { if (mol.biol=="ALL1/AF4") 1 else 2 }
patientcolors <- unlist(lapply(esetSel$mol.bio, color.map))
hc<-hclust(dist(t(data)))
dd.col<-as.dendrogram(hc)
groups <- cutree(hc,k=7)
annotation<-data.frame(Var1=factor(patientcolors,labels=c("class1","class2")),Var2=groups)
pheatmap(data, annotation=annotation, fontsize=9, fontsize_row=6) #为样品分组

Var1 = c("navy", "skyblue")
Var2 = c("snow", "steelblue")
names(Var1) = c("class1", "class2")
ann_colors = list(Var1 = Var1, Var2 = Var2)
pheatmap(data, annotation=annotation, annotation_colors = ann_colors, fontsize=9, fontsize_row=6) #为分组的样品设定颜色